EV Battery Charging Schedule Using ML for SoC and SoH Balance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing battery management systems in electric vehicles do not effectively optimize the state of health (SoH) of the battery by controlling charging patterns, leading to reduced battery lifespan and efficiency.
Innovation Solution
A machine learning-based system that utilizes a regional model to generate weights for a local model, optimizing battery charging schedules by considering driver behavior, vehicle utilization, and regional data such as weather and traffic conditions, maintaining the battery state of charge (SoC) within a desired range to enhance SoH.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If the battery is charged to 100% capacity, then the energy storage is maximized, but the battery lifespan is reduced
Solution Approach 1:
The charging system dynamically adjusts the target state of charge based on real-time battery conditions, vehicle usage patterns, and environmental factors. Instead of a fixed 100% charging target, the system adapts the charging level to optimize both energy storage and battery longevity, maintaining charge between 20-80% under normal conditions while allowing 100% charge only when necessary.
Solution Approach 2:
The system changes the charging parameter (state of charge target) based on multiple factors including battery temperature, age, usage patterns, and climate conditions. By varying the target charge level dynamically rather than always charging to 100%, the system extends battery life while still meeting energy storage needs.
2Loss of time
If fast charging is used, then the charging time is reduced, but the battery wear is increased
Solution Approach 1:
The system employs periodic charging cycles that alternate between fast charging and slower charging phases. During fast charging, the system monitors battery temperature and charge level closely, then inserts slower charging periods to allow thermal management and reduce stress on battery cells. This periodic alternation enables faster overall charging while mitigating battery wear.
Solution Approach 2:
Before initiating fast charging, the system performs preliminary assessments of battery conditions including temperature, state of charge, and recent usage patterns. If conditions are not optimal for fast charging, the system prepares the battery through preliminary conditioning (such as thermal management or slower initial charging) to create safe conditions for subsequent fast charging, thereby reducing wear.
3Duration of action of stationary object
If the charging schedule is customized using machine learning, then the battery life is extended, but the system complexity is increased
Solution Approach 1:
The charging system uses machine learning models that automatically learn and adapt to user charging patterns, vehicle usage, and environmental conditions without requiring manual programming or complex configuration. The system serves itself by autonomously optimizing charging schedules based on data it collects, reducing the need for complex external control mechanisms while extending battery life.
Solution Approach 2:
The system implements continuous feedback loops where charging outcomes are monitored and fed back to the machine learning model. This feedback mechanism allows the system to automatically refine its charging strategies based on actual battery performance and user behavior, extending battery life through adaptive optimization without requiring complex manual intervention or system reconfiguration.
Data Source
AI summary
Controlling a battery state of charge using machine learning is provided. A system of an electric vehicle identifies data including a state of charge of a battery of the electric vehicle, a state of health of the battery of the electric vehicle, and a drive mode of the electric vehicle. The system establishes, based on input of the data into a local model configured on the electric vehicle and trained with machine learning, a schedule to control charging of the battery of the electric vehicle. The system executes, responsive to a power source electrically coupled to the battery of the electric vehicle, the schedule to control an amount of current supplied from the power source to the battery of the electric vehicle.